Updated: Aug 26
Knowledge Hub › Technical Insights
TECHNICAL INSIGHTS
How ATP Thermal Performance Testing Works: From Test Conditions to Technical Evidence
A practical engineering guide to ATP thermal performance testing, covering test configuration, controlled conditions, stabilization, instrumentation, data quality and the evidence chain behind technically defensible results.
22 August 2026 · Technical Insights · 10 min read
KEY TAKEAWAY
A defensible ATP thermal-performance result depends on a complete evidence chain linking the tested configuration, controlled conditions, stabilization, instrumentation, data quality, analysis and final technical evidence.
IN THIS GUIDE
ATP thermal performance testing is sometimes reduced to a simple question: Did the equipment reach the required temperature? From an engineering perspective, that is not enough.
A technically defensible ATP result depends on a complete evidence chain linking the equipment configuration, applicable test method, controlled environmental conditions, instrumentation, stabilization, recorded data, calculations and acceptance criteria. The test result is therefore not just the final number. It is the output of a controlled measurement process.
Regulatory status
This guide has been prepared against the ATP Agreement applicable from 22 June 2024 and cross-checked against the ATP Handbook incorporating amendments that enter into force on 25 August 2026. Until that date, the 22 June 2024 text remains the applicable edition.
1. Thermal performance testing is an evidence process, not only a test run
A good ATP test programme begins before the equipment enters the test chamber.
A useful engineering sequence is: Requirement → Configuration → Test Method → Controlled Conditions → Measurement → Stabilization → Analysis → Technical Evidence.
Each stage affects the validity of the next. A perfectly executed test on an incorrectly defined configuration can produce technically irrelevant evidence. Likewise, a correctly configured body tested under poorly controlled conditions may produce data that cannot reliably support the intended conclusion.
Test preparation should begin by defining exactly what is being demonstrated, on which configuration and against which ATP provision.
2. What ATP thermal performance testing is intended to demonstrate
ATP contains different technical verification routes, and they should not be treated as one universal ‘ATP test’. At a high level, the thermal evidence may concern:
Insulating capacity — the overall heat-transfer performance of the insulated body, commonly expressed through the K coefficient.
Thermal appliance effectiveness — the capability of refrigerated, mechanically refrigerated, heated or mechanically refrigerated-and-heated equipment to establish and maintain the temperature conditions associated with its class.
In-service or periodic verification — procedures used within the ATP framework to verify continued conformity of equipment already in service.
Are we demonstrating the thermal performance of the insulated body, the capability of the thermal appliance, or the continued performance of an existing equipment configuration?
These are related questions, but they are not interchangeable.
3. Define the test configuration before generating evidence
The test article should be treated as a controlled configuration. For the insulated body, relevant configuration characteristics may include:
principal dimensions and internal volume
wall, floor and roof construction
insulation system
doors and seals
structural members
penetrations and openings
accessories capable of affecting heat transfer
Where a thermal appliance is installed, the evidence chain may additionally depend on:
appliance type
compressor or other drive arrangement
refrigerant configuration
heat exchangers
evaporator and condenser arrangements
fans and airflow distribution
control system
installation interfaces
The objective is not to document every component for administrative completeness. It is to ensure that the physical equipment being assessed can be traced to the technical evidence generated during testing.
A later configuration change should trigger the question: Could this change affect the thermal basis demonstrated by the test?
4. Controlled test conditions are part of the measurement
For K-coefficient testing of equipment other than liquid-foodstuff tanks, ATP requires the empty body to be placed in an insulated chamber and allows the coefficient to be measured using either internal cooling or internal heating.
The prescribed conditions illustrate why thermal testing cannot be separated from the test environment. For example, the ATP method specifies a temperature difference between the inside of the body and the insulated chamber of 25 °C ± 2 °C, while maintaining the average wall temperature at +20 °C ± 0.5 °C.
Air around the outside of the body is continuously circulated, with air velocity measured 10 cm from the walls maintained between 1 and 2 m/s. For the internal heating or cooling arrangement, sufficient air circulation must also be provided to establish acceptable temperature uniformity.
The method further controls temperature distribution: once continuous operation has been established, the temperature spread between specified measurement locations must remain within defined limits.
The test chamber is part of the measurement system.
Poor chamber uniformity, uncontrolled airflow or an unsuitable internal air-distribution arrangement can influence the measured heat transfer even when the body itself has not changed.
5. Stabilization is not the same as reaching a temperature
One of the most important distinctions in thermal testing is the difference between reaching the target condition and establishing a stable condition suitable for calculation.
For K-coefficient testing, ATP defines specific steady-state requirements. The mean inside and outside temperatures are evaluated over a steady-state period of at least 12 hours. Their variation during that period must remain within defined limits, and additional criteria are used to confirm that the thermal input or cooling capacity has also stabilized.
For the K calculation, mean values over at least the final six hours of the steady-state period are used. The ATP provisions also limit the allowable difference between mean temperatures at the beginning and end of that calculation period.
A flat-looking temperature trace is not automatically evidence of steady state.
Stability should be demonstrated using the criteria applicable to the test method rather than selected visually because the data ‘looks stable’.
Engineering practice versus ATP requirement
Higher-frequency data logging, automatic stability calculations and graphical trend evaluation can significantly improve test control and diagnostics. However, these should be described correctly. They are engineering tools that can support the ATP method; they should not be presented as ATP requirements unless the Agreement explicitly requires them.
6. Sensor location is part of the test definition
A temperature value has limited meaning without knowing where it was measured.
For parallelepipedic equipment, ATP defines the mean inside temperature using measurements taken 10 cm from the walls at 12 locations: the eight inside corners and the centres of the four inside faces having the largest area. The corresponding mean outside temperature is determined using an equivalent arrangement outside the body. Temperature measuring instruments are to be protected against radiation.
ATP evaluates the thermal state of the body using a defined spatial measurement system, not a single convenient temperature sensor.
Sensor mapping should therefore be documented before the test, not reconstructed afterward.
7. Measurement quality belongs inside the evidence chain
A test report can contain hundreds of data points and still provide weak evidence if the measurement system is not technically controlled. A robust test strategy should address:
Calibration — Are the instruments calibrated over the relevant measurement range?
Traceability — Can the measurement results be related to a documented calibration chain?
Sensor identification — Can each data channel be traced to a physical measurement location?
Resolution and accuracy — Are they adequate relative to the acceptance criteria being evaluated?
Data acquisition integrity — Are timestamps, channel mapping and acquisition continuity controlled?
Corrections — Are corrections such as electrical cable losses or other method-specific adjustments identified and traceable?
ATP itself contains accuracy provisions for K-coefficient determination and requires specific treatment of quantities contributing to the calculated heat transfer.
Engineering practice
Recording data more frequently than the formal minimum can be valuable for identifying transient behaviour, detecting cycling, evaluating defrost events, checking control stability, detecting sensor anomalies and defining the valid calculation window. The additional data improves engineering visibility; it does not change the applicable ATP acceptance criterion.
8. K-coefficient testing and thermal appliance testing answer different questions
The K coefficient describes the overall heat-transfer performance of the complete insulated body.
Conceptually: K = W / (S × ΔT)
The result relates the heat-transfer rate to the mean heat-transfer surface and the temperature difference across the body. It is therefore a system-level property. A refrigeration appliance capacity value does not determine K, and a good K coefficient does not by itself demonstrate that a refrigeration appliance can satisfy the required thermal class.
Mechanically refrigerated equipment
For mechanically refrigerated equipment, ATP uses a different test logic. The equipment is evaluated in an insulated chamber, with the external environment associated with the class test conditions. The appliance must lower the temperature inside the empty body and then demonstrate the capability associated with the relevant class.
Classes A, B and C use selectable practically constant internal temperature ranges, while Classes D, E and F use fixed practically constant maximum temperatures. For Classes B, C, E and F, the K coefficient must in every case be no greater than 0.40 W/m²·K.
K testing asks how readily heat passes through the body. Thermal-appliance testing asks whether the configured equipment can produce and maintain the required thermal condition.
Both may be necessary to support the classification, but they demonstrate different characteristics.
9. Why test conditions can change the interpretation
Air distribution — Poor air circulation can produce local hot or cold regions that distort the relationship between the measured mean temperature and the actual thermal behaviour of the body.
Sensor radiation effects — A sensor influenced by radiant heat rather than representative air temperature can introduce systematic bias. This is why ATP specifies radiation-protected temperature measurements.
Thermal transients — Using data before the required stable condition has been demonstrated can cause stored thermal energy in the body to be interpreted incorrectly as steady heat transfer.
Appliance cycling and defrosting — Control cycles and defrost events may be normal equipment behaviour, but their treatment must follow the applicable test procedure rather than being selectively removed because they make the data less convenient.
Electrical losses — In K testing using internal heating, measured electrical input is not automatically identical to useful heat released inside the test body. ATP specifically addresses cable losses between the measuring instrument and the tested body.
A defensible result requires more than collecting values from sensors. It requires understanding what each measurement physically represents.
10. From raw measurements to defensible technical evidence
A professional ATP test evidence chain should be reproducible. A useful engineering sequence is:
RAW DATA → CHANNEL & SENSOR VALIDATION → TEST-CONDITION CHECK → STABILIZATION / VALID DATA WINDOW → METHOD-SPECIFIC CORRECTIONS → AVERAGED TEST QUANTITIES → CALCULATION OR PERFORMANCE ASSESSMENT → ACCEPTANCE CRITERIA → TRACEABLE TEST RESULT
For K testing, ATP uses the mean values from the defined stable calculation period rather than an arbitrary snapshot. This distinction becomes especially important when the result is close to a classification limit.
A result should not merely be numerically below a threshold. The measurement process supporting that result should also be technically defensible.
11. Common ATP thermal-testing mistakes
Mistake 1 — Starting the test before the configuration is frozen. If the tested configuration cannot later be reconstructed, the value of the evidence is weakened.
Mistake 2 — Declaring steady state because the graph looks flat. Use the applicable stability criteria rather than visual judgement alone.
Mistake 3 — Averaging transient data. A large data set does not compensate for using the wrong calculation window.
Mistake 4 — Treating refrigeration-unit nameplate capacity as demonstrated equipment performance. Installed-system performance and catalogue capacity are not the same technical statement.
Mistake 5 — Failing to document sensor positions. Temperature measurements must remain traceable to their physical locations.
Mistake 6 — Treating calibration as an administrative certificate. Calibration affects confidence in the result and should be reviewed in relation to the measurement range and acceptance criteria.
Mistake 7 — Applying undocumented data corrections after the test. Corrections should arise from the test method or a controlled engineering rationale and remain traceable.
Mistake 8 — Treating a technically successful test as automatic certification. ATP certification and approval remain within the framework of the relevant Contracting Party and competent authority.
12. ATP thermal test readiness checklist
Before formal thermal performance testing begins, the engineering team should be able to answer:
What ATP requirement or technical characteristic is the test intended to demonstrate?
Is the test configuration clearly defined and frozen?
Is the selected ATP test method applicable to that configuration?
Can the required chamber temperature and airflow conditions be controlled?
Are all required temperature measurement locations defined?
Are the instruments suitable, calibrated and traceable?
Are data acquisition channels and sensor identities controlled?
Is the applicable stabilization criterion defined before the test starts?
Are calculations, corrections and acceptance criteria documented in advance?
Will the final test report allow the configuration, measurements and result to be reconstructed later?
If several of these questions cannot yet be answered, the project may have a test-readiness gap before it has a test-result problem.
13. The engineering perspective
ATP thermal performance testing is most effective when treated as a controlled engineering measurement rather than a chamber activity performed at the end of development.
A strong test strategy should establish, before testing, what will be demonstrated; which configuration represents the equipment; which conditions must be controlled; what will be measured; how stabilization will be demonstrated; which data will be used; how the result will be calculated; and which evidence will support the final conclusion.
When these elements are defined in advance, testing becomes more repeatable, interpretation becomes clearer and avoidable retesting becomes less likely.
Testing should generate evidence against a defined technical question — not generate data first and search for an interpretation afterward.
Primary references
UNECE — Agreement on the International Carriage of Perishable Foodstuffs and on the Special Equipment to be Used for such Carriage (ATP), version applicable from 22 June 2024.
UNECE — ATP Handbook incorporating amendments entering into force on 25 August 2026.
The ATP Handbook contains explanatory comments intended to support interpretation, harmonization and application; those comments are not themselves legally binding on Contracting Parties.
ATP Compliance is an independent technical information and advisory platform and is not affiliated with UNECE, national competent authorities or official ATP test stations.
Related Technical Guides
Need technical guidance?
ATP Compliance provides independent engineering guidance on ATP requirements, testing strategy, measurement quality and compliance readiness.
Comments